Authors:
David Govi
;
Alessandro Rizzuto
;
Federico Schipani
and
Alessandro Lazzeri
Affiliation:
Deepclever S.r.l., Via Bure Vecchia Nord n.c. 115, 51100, Pistoia (PT), Italy
Keyword(s):
Two-stage Genetic Algorithm, Working Centers, Flexible Job Shop Scheduling, Chromosome Representation, Local Search, Population Initialization.
Abstract:
Inspired by industrial issues and demands, we define a novel version of the Flexible Job Shop Scheduling Problem with Working Center. A working center is a group of machines performing the same type of operation. The job operations of different types follow a strict sequence across the working centers, while any order is allowed among operations of the same type. This paper illustrates a genetic algorithm with a two-stage chromosome representation, adapted genetic operators, local search, and social disaster technique to deal with a real-world industrial application. The algorithm has been tested on a classical benchmark to assess its adaptability and compare its performance with state-of-the-art techniques; then, we tested different variations of the proposed algorithm on a real-case test instance showing a consistent improvement when compared with the heuristic in use at the industrial company.